edu.illinois.cs.cogcomp.lbjava.learn

Class NaiveBayes.NaiveBayesVector

java.lang.Object

edu.illinois.cs.cogcomp.lbjava.learn.SparseWeightVector

edu.illinois.cs.cogcomp.lbjava.learn.NaiveBayes.NaiveBayesVector

All Implemented Interfaces:

Serializable, Cloneable

Enclosing class:

NaiveBayes

protected class NaiveBayes.NaiveBayesVector
extends SparseWeightVector

Keeps track of all the counts associated with a given label. Features are associated with NaiveBayes.Counts. Those not appearing in this vector are assumed to have a count of 0. The invocation of either of the scaledAdd methods increments the prior count for the label.

RealFeatures' strengths are ignored by this vector; they are assumed to be equal to 1, as if the feature were an active Boolean feature.

Author:

Nick Rizzolo

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected edu.illinois.cs.cogcomp.core.datastructures.vectors.OVector	counts The counts in the vector indexed by their Lexicon key.
protected NaiveBayes.Count	priorCount The prior count is the number of times either scaledAdd method has been called.

Fields inherited from class edu.illinois.cs.cogcomp.lbjava.learn.SparseWeightVector

defaultCapacity, defaultWeight, weights

Constructor Summary

Constructors

Constructor and Description
NaiveBayesVector() Simply instantiates counts.
NaiveBayesVector(NaiveBayes.Count[] w) Simply initializes counts.
NaiveBayesVector(edu.illinois.cs.cogcomp.core.datastructures.vectors.OVector w) Simply initializes counts.

Method Summary

Modifier and Type	Method and Description
Object	clone() Returns a copy of this NaiveBayesVector.
double	dot(int[] exampleFeatures, double[] exampleValues) Takes the dot product of this vector with the given vector, using the hard coded smoothing weight.
double	dot(int[] exampleFeatures, double[] exampleValues, double defaultW) Takes the dot product of this vector with the given vector, using the specified default weight when encountering a feature that is not yet present in this vector.
SparseWeightVector	emptyClone() Returns a new, empty weight vector with the same parameter settings as this one.
double	getCount(int featureIndex) Returns the count of the given feature.
NaiveBayes.Count	getPrior() Returns the prior count of the prediction value associated with this vector.
double	getWeight(int featureIndex, double defaultW) Returns the weight of the given feature
void	incrementCount(int featureIndex, double factor) Increments the count of the given feature.
void	read(edu.illinois.cs.cogcomp.core.datastructures.vectors.ExceptionlessInputStream in) Reads the representation of a weight vector with this object's run-time type from the given stream, overwriting the data in this object.
void	scaledAdd(int[] exampleFeatures, double[] exampleValues, double factor) This method is similar to the implementation in SparseWeightVector except that incrementCount(int,double) is called instead of SparseWeightVector.setWeight(int,double).
void	scaledAdd(int[] exampleFeatures, double[] exampleValues, double factor, double defaultW) This method is similar to the implementation in SparseWeightVector except that the defaultW argument is ignored and incrementCount(int,double) is called instead of SparseWeightVector.setWeight(int,double).
void	setWeight(int f, double w) This method is overridden to do nothing; use incrementCount(int,double) instead.
void	write(edu.illinois.cs.cogcomp.core.datastructures.vectors.ExceptionlessOutputStream out) Writes the weight vector's internal representation in binary form.
void	write(PrintStream out) Outputs the contents of this vector into the specified PrintStream.
void	write(PrintStream out, Lexicon lex) Outputs the contents of this vector into the specified PrintStream.

Methods inherited from class edu.illinois.cs.cogcomp.lbjava.learn.SparseWeightVector

clear, getWeight, pairwiseMultiply, readWeightVector, scaledAdd, scaledMultiply, scaledMultiply, setWeight, size, toString, toString, toStringJustWeights, toStringJustWeights

Methods inherited from class java.lang.Object

equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

counts

protected edu.illinois.cs.cogcomp.core.datastructures.vectors.OVector counts

The counts in the vector indexed by their Lexicon key.

priorCount

protected NaiveBayes.Count priorCount

The prior count is the number of times either scaledAdd method has been called.

Constructor Detail

NaiveBayesVector

public NaiveBayesVector()

Simply instantiates counts.

NaiveBayesVector

public NaiveBayesVector(NaiveBayes.Count[] w)

Simply initializes counts.

Parameters:

w - An array of counts.

NaiveBayesVector

public NaiveBayesVector(edu.illinois.cs.cogcomp.core.datastructures.vectors.OVector w)

Simply initializes counts.

Parameters:

w - A vector of counts.

Method Detail

getPrior

public NaiveBayes.Count getPrior()

Returns the prior count of the prediction value associated with this vector.

dot

public double dot(int[] exampleFeatures,
                  double[] exampleValues)

Takes the dot product of this vector with the given vector, using the hard coded smoothing weight.

Overrides:

dot in class SparseWeightVector

Parameters:

exampleFeatures - The example's array of feature indices.

exampleValues - The example's array of feature values.

Returns:

The computed dot product.

dot

public double dot(int[] exampleFeatures,
                  double[] exampleValues,
                  double defaultW)

Takes the dot product of this vector with the given vector, using the specified default weight when encountering a feature that is not yet present in this vector. Here, weights are taken as log(feature count / prior count). The output of this method is related to the empirical probability of the example e as follows:

exp(dot(e)) / (sum of all labels' prior counts)) =
P(e's label && e)

Overrides:

dot in class SparseWeightVector

Parameters:

exampleFeatures - The example's array of feature indices.

exampleValues - The example's array of feature values.

defaultW - The default weight.

Returns:

The computed dot product.

getCount

public double getCount(int featureIndex)

Returns the count of the given feature.

Parameters:

featureIndex - The feature index.

Returns:

The count of the feature.

getWeight

public double getWeight(int featureIndex,
                        double defaultW)

Returns the weight of the given feature

Overrides:

getWeight in class SparseWeightVector

Parameters:

featureIndex - The feature index.

defaultW - The default count.

Returns:

The weight of the feature.

setWeight

public void setWeight(int f,
                      double w)

This method is overridden to do nothing; use incrementCount(int,double) instead.

Overrides:

setWeight in class SparseWeightVector

Parameters:

f - Unused.

w - Unused.

incrementCount

public void incrementCount(int featureIndex,
                           double factor)

Increments the count of the given feature.

Parameters:

featureIndex - The index of the feature to update.

factor - The factor by which to increment.

scaledAdd

public void scaledAdd(int[] exampleFeatures,
                      double[] exampleValues,
                      double factor)

This method is similar to the implementation in SparseWeightVector except that incrementCount(int,double) is called instead of SparseWeightVector.setWeight(int,double).

Overrides:

scaledAdd in class SparseWeightVector

Parameters:

exampleFeatures - The example's array of feature indices.

exampleValues - The example's array of feature values.

factor - The scaling factor.

scaledAdd

public void scaledAdd(int[] exampleFeatures,
                      double[] exampleValues,
                      double factor,
                      double defaultW)

This method is similar to the implementation in SparseWeightVector except that the defaultW argument is ignored and incrementCount(int,double) is called instead of SparseWeightVector.setWeight(int,double).

Overrides:

scaledAdd in class SparseWeightVector

Parameters:

exampleFeatures - The example's array of feature indices.

exampleValues - The example's array of feature values.

factor - The scaling factor.

defaultW - Unused.

write

public void write(PrintStream out)

Outputs the contents of this vector into the specified PrintStream. The string representation is the same as in the super class, except the "Begin" annotation line also contains the value of priorCount in parentheses. In addition, this method has access to the lexicon, so the output of this method is equivalent to that of write(PrintStream,Lexicon).

Overrides:

write in class SparseWeightVector

Parameters:

out - The stream to write to.

write

public void write(PrintStream out,
                  Lexicon lex)

Outputs the contents of this vector into the specified PrintStream. The string representation is the same as in the super class, except the "Begin" annotation line also contains the value of priorCount in parentheses.

Overrides:

write in class SparseWeightVector

Parameters:

out - The stream to write to.

lex - The feature lexicon.

write

public void write(edu.illinois.cs.cogcomp.core.datastructures.vectors.ExceptionlessOutputStream out)

Writes the weight vector's internal representation in binary form. Note: this method does not call SparseWeightVector.write(ExceptionlessOutputStream) and does not output its class name or the contents of SparseWeightVector.weights since there shouldn't be any.

Overrides:

write in class SparseWeightVector

Parameters:

out - The output stream.

read

public void read(edu.illinois.cs.cogcomp.core.datastructures.vectors.ExceptionlessInputStream in)

Reads the representation of a weight vector with this object's run-time type from the given stream, overwriting the data in this object.

This method is appropriate for reading weight vectors as written by write(ExceptionlessOutputStream).

Overrides:

read in class SparseWeightVector

Parameters:

in - The input stream.

clone

public Object clone()

Returns a copy of this NaiveBayesVector.

Overrides:

clone in class SparseWeightVector

Returns:

A copy of this NaiveBayesVector.

emptyClone

public SparseWeightVector emptyClone()

Returns a new, empty weight vector with the same parameter settings as this one.

Overrides:

emptyClone in class SparseWeightVector

Returns:

An empty weight vector.